Volume 122, Issue 4, Pages 619-631 (August 2005) Interaction with Vesicle Luminal Protachykinin Regulates Surface Expression of δ- Opioid Receptors and Opioid Analgesia  Ji-Song Guan, Zhen-Zhong Xu, Hua Gao, Shao-Qiu He, Guo-Qiang Ma, Tao Sun, Li-Hua Wang, Zhen-Ning Zhang, Isabelle Lena, Ian Kitchen, Robert Elde, Andreas Zimmer, Cheng He, Gang Pei, Lan Bao, Xu Zhang  Cell  Volume 122, Issue 4, Pages 619-631 (August 2005) DOI: 10.1016/j.cell.2005.06.010 Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 1 Requirement of Preprotachykinin A Expression for LDCV Localization of DOR in Small DRG Neurons (A) Double-immunofluorescence labeling shows colocalization of DOR (in green) and substance P (SP, in red) in LDCVs in small DRG neurons of the mouse. Pre-embedding immunogold-silver labeling of DOR combined with postembedding immunogold (10 nm in diameter) labeling of substance P shows localization of DOR (arrowhead) in the membrane of LDCV and substance P (arrow) in the lumen of LDCV. Quantitative analysis at the ultrastructural level shows that the LDCVs carrying DOR almost exclusively contain substance P (a total of 180 LDCVs in three experiments). Data are represented as mean ± SEM. (B) Double-immunofluorescence labeling shows that DOR (in green) and CGRP (in red) are colocalized in LDCVs in small DRG neurons of wild-type (+/+) mice but not in LDCVs of PPT-A knockout (−/−) mice. The number of CGRP-containing vesicles in small DRG neurons of PPT-A knockout mice is unchanged (for both +/+ and −/− neurons, ∼7 LDCVs/μm2 in the optic section through the center of neuron). (C) In small DRG neurons of PPT-A knockout mice, immunogold-silver labeling of DOR is seen in multivesicular bodies (arrowhead) instead of in LDCVs (mean ± SEM, n = 18 neuron profiles). **p < 0.01 versus the neurons of wild-type mice. Postembedding immunogold labeling shows that, in small DRG neurons of PPT-A knockout mice, CGRP is still localized in LDCVs (arrowhead). Scale bars, 5 μm for confocal images; 100 nm for electron micrographs. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 2 Essential Role of Protachykinin in Biogenesis of DOR-Containing LDCV In HEK293 cells, vectors expressing HA-tagged DOR (HA-DOR), chromogranin A (CgA), and wild-type or mutated β-preprotachykinin (β-PPT) or α-CGRP are transfected in various combinations (A–G). Transfected cells are immunofluorescence stained and examined in confocal microscope. HA-DOR is labeled with HA antibody. Scale bar, 10 μm for (A), (B), and (D)–(F). (A) In cells expressing HA-DOR, DOR is localized on the cell surface and in the cytoplasm. (B) In the cell expressing chromogranin A-GFP and HA-DOR, DOR (in red) is not localized in chromogranin A (in green) vesicles (arrowhead). (C) Diagram showing wild-type (WT) or mutants (M) of β-PPT and α-CGRP proteins. SP, substance P; SK, substance K. (D) In the cell expressing chromogranin A-GFP and β-PPT(1–68) (β-PPT M1) with the domains of signal peptide, N-terminal propeptide, and substance P, substance P (in red) is localized in chromogranin A (in green) vesicles (arrowhead). Substance P is labeled with its antibody. (E) In the cell expressing HA-DOR, chromogranin A-GFP, and β-PPT(1–68), DORs (in red), substance P (in blue), and chromogranin A (in green) are colocalized in vesicles (arrowhead, in white) (see Figure S2A for images of separated channels and the analysis of colocalization). (F) In the cell expressing HA-DOR, chromogranin A tagged with c-Myc epitope, and β-PPT(1–57, 69–130)-GFP (β-PPT M2) missing the substance P domain, DOR (in red) is not localized in β-PPT(1–57, 69–130) (in green) and chromogranin A (in blue) containing vesicles (arrowhead, in cyan) (see Figure S2C for images of separated channels). Chromogranin A-Myc is labeled with c-Myc antibody. (G) Quantitative analysis shows that, in HEK293 cells coexpressing DOR, chromogranin A, and β-PPT(1–130) or β-PPT(1–68) or α-CGRP(1–82)/β-PPT(58–68) (α-CGRP M2), DOR is localized in chromogranin A vesicles. **p < 0.01 versus the cell coexpressing DOR and chromogranin A (mean ± SEM, n = 30 cells/group). (H) Pre-embedding immunogold-silver labeling with HA antibody shows that, in cells triple-transfected with vectors expressing HA-DOR, β-PPT(1–68), and chromogranin A-GFP, DORs (arrowhead) associate with the membrane of LDCVs. **p < 0.01 versus the cells transfected with HA-DOR (mean ± SEM, 12 cells/group). n.d., not detected. Scale bar, 100 nm. (I) Double-immunofluorescence labeling shows that, in a cultured small DRG neuron of the newborn PPT-A knockout (−/−) mice, the colocalization of DOR and CGRP in LDCVs (arrows) is rescued by the transfection with the vector expressing β-PPT(1–68)-GFP. **p < 0.01 (mean ± SEM, n = 12 neuron profiles/group). Scale bar, 5 μm. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 3 Interaction between DOR and Substance P Domain of Protachykinin (A) Coimmunoprecipitation (IP) and immunoblotting (IB) show that, in HEK293 cells cotransfected with vectors expressing β-PPT(58–68)-GFP and HA-DOR, β-PPT(58–68)-GFP (∼30 kDa) is found in HA-antibody-precipitated proteins. This interaction is inhibited by adding 1 μM substance P into the lysate. In cells cotransfected with vectors expressing HA-DOR and α-CGRP(83–119)-GFP (∼33 kDa), no significant CoIP signal of CGRP-GFP is detected. **p < 0.01, lanes 3 and 4 versus lanes 1 and/or 2 (mean ± SEM, n = 3). (B) In HEK293 cells cotransfected with vectors expressing β-PPT(58–68)-GFP and HA-DOR, HA-DOR (∼60 kDa) is identified in GFP-antibody-precipitated proteins. **p < 0.01, lane 2 versus lane 1 (mean ± SEM, n = 3). (C) Double-immunofluorescence labeling shows colocalization of DOR and substance P in LDCVs in PC12 cells. After crosslinking with membrane-permeable disuccinimidyl suberate, in DOR-antibody-precipitated proteins, the substance P-immunoreactive band is seen at the same position where the DOR band (∼60 kDa) is found, while CGRP is not detected. In substance P-antibody-precipitated proteins, DOR is also detected. The data represent three independent experiments. Scale bar, 5 μm. (D) After crosslinking with membrane-permeable disuccinimidyl suberate in dissociated mouse DRG neurons, immunoblot of substance P is seen in DOR-antibody-precipitated proteins. The data represent three independent experiments. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 4 Identification of the Sorting Domain of DOR (A) Diagram shows wild-type (WT) and mutants (M) of DOR and MOR and summarizes the results of the localization of these molecules tagged with HA in transfected PC12 cells based on immunofluorescence labeling with HA antibody. Domain swap between DOR and MOR (DOR M1–M3; MOR M1) shows that the third luminal (extracellular) domain of DOR (285–293) is responsible for the LDCV localization. Point mutation of DOR (DOR M4–M7) indicates that D288 and D293 are critical for LDCV localization. TM, transmembrane domain; −, not in LDCVs; ++, mainly in LDCVs; +, in some LDCVs. (B) Representative confocal images show the LDCV localization of the DOR mutant with DOR(1–262) replaced by MOR(1–282) (DOR M1). The DOR with point mutation at D288 and D293 (DOR M4) is not localized in vesicles. Scale bar, 5 μm. (C) Pre-embedding immunogold-silver labeling with HA antibody in transfected PC12 cells shows that DOR, but not DOR(D288A, D293A), associates with the membrane of LDCVs. Quantitative analysis shows that the immunogold-silver labeling on LDCVs is decreased in cells expressing the DOR mutant with DOR(263–372) replaced by MOR(283-398) (DOR M2) or DOR(D288A, D293A) (DOR M4) but not with DOR(R291A, R292A) (DOR M7). The decrease is less pronounced in cells expressing DOR(D288A) (DOR M5). The increase in labeling in multivesicular bodies is seen in cells expressing DOR M2 and M4. **p < 0.01 versus DOR (mean ± SEM, n = 25 cell profiles/group). Scale bars, 100 nm. (D) In contrast to the cell-surface labeling of MOR, the mutated MOR with the third luminal (extracellular) domain of DOR (MOR M1) associates with vesicles. The MOR mutant with the point-mutated third luminal domain (D288A, D293A) of DOR (MOR M2) is not localized in vesicles. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 5 Interaction of the Sorting Domain of DOR with the Substance P Domain of Protachykinin (A) Coimmunoprecipitation (IP) and immunoblotting (IB) show that the immunoblot signal of substance P [β-PPT(58–68)]-GFP (∼30 kDa) is decreased in HA-antibody-precipitated proteins from HEK293 cells cotransfected with vectors expressing β-PPT(58–68)-GFP and HA-DOR(D288A, D293A) (DOR M4). Neither HA-DOR mutant with the C terminus of MOR (DOR M3) nor DOR(R291A, R292A) (DOR M7) attenuates the immunoblot signal of β-PPT(58–68)-GFP. **p < 0.01 versus HA-DOR (mean ± SEM, n = 3). (B) The immunoblot signal of β-PPT(58–68)(R58A)-GFP is weaker than that of β-PPT(58–68)-GFP in Myc-antibody-precipitated proteins from HEK293 cells cotransfected with vectors expressing c-Myc-epitope-tagged DOR (DOR-Myc) and the peptide-GFP. β-PPT(58–68)(R58A, K60A) produces a more pronounced reduction in the signal. **p < 0.01 versus β-PPT(58–68) (mean ± SEM, n = 3). (C) After crosslinking with membrane-permeable disuccinimidyl suberate in PC12 cells transfected with the vector expressing HA-DOR, substance P immunoblot is seen at the same position as DOR (∼60 kDa) in HA-antibody-precipitated proteins. The signal of substance P is decreased in cells expressing HA-DOR(D288A, D293A) (DOR M4) or HA-DOR mutant with DOR(263–372) replaced by MOR(283–398) (DOR M2). **p < 0.01 versus HA-DOR (mean ± SEM, n = 3). (D) [3H]substance P binds to GST-fused DOR(281–301), which corresponds roughly to the third luminal (extracellular) domain, but not to GST-fused DOR(105–123) and DOR(184–213), which correspond roughly to the first and second luminal domain, respectively. [3H]substance P is not bound to the third luminal domain of DOR with mutation at D288 and D293. GST protein serves as control. **p < 0.01 versus GST (mean ± SEM, n = 9). Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 6 Attenuation of DOR Distribution in Small DRG Neurons of PPT-A Knockout Mice (A) Dissociated DRG neurons from wild-type (+/+) and PPT-A knockout (−/−) mice are treated with 1 μM deltorphin I (Delt) or 1 μM capsaicin for 10 min. Cells are fixed, permeabilized, and immunofluorescence labeled. Confocal images show that deltorphin induces distinct surface-associated labeling in small DRG neurons (arrow), while the number of intracellular vesicles with the receptor is reduced. The same treatment does not change DOR distribution in small DRG neurons from PPT-A knockout mice. For semiquantification, the number of small neurons exhibiting distinct surface-associated DOR labeling in the surface zone (1.5 μm in width) (Bao et al., 2003) (arrow-pointed neurons) is counted. **p < 0.01 versus control (mean ± SEM, n = 200 neuron profiles/group). Scale bar, 5 μm. (B) Cell-surface biotinylation/immunoblotting shows that 10 min treatment of 1 μM deltorphin or capsaicin increases surface insertion of DOR in DRG neurons from wild-type mice but not in DRG neurons from PPT-A knockout mice. *p < 0.05, DOR signal relative to actin of treated group versus that of control (mean ± SEM, n = 3). (C) Small DRG neurons give rise to C fibers. Peripheral terminals of C fibers in the skin are activated by nociceptive stimulation, leading to the release of neurotransmitters at the central terminii in laminae I–II of the spinal cord. (D) Immunoblotting shows that the DOR level is reduced in the dorsal horn of the spinal cord of PPT-A knockout mice. *p < 0.05 versus the wild-type (mean ± SEM, n = 3). (E) Immunofluorescence labeling of DOR is decreased in laminae I–II of the spinal cord of PPT-A knockout mice. Scale bar, 50 μm. (F) Quantitative autoradiographic analysis of [3H]deltorphin I binding shows that DOR binding sites in laminae I–II of the spinal cord of PPT-A knockout mice is significantly decreased compared with that in the spinal cord of wild-type mice. **p < 0.01 versus the wild-type (mean ± SEM, n = 3). (G) Pre-embedding immunogold-silver labeling shows that, in the axonal terminals in lamina II of the spinal cord of PPT-A knockout mice, DOR is seen in endosome-like vesicles (arrowhead), while LDCVs (arrows) are not labeled. **p < 0.01 versus the wild-type (mean ± SEM, n = 18 terminals). Scale bar, 100 nm. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 7 Loss of DOR-Mediated Spinal Analgesia and Morphine Tolerance in PPT-A Knockout Mice (A) In the 52°C water tail-immersion test, no significant difference of the tail-flick latency between PPT-A knockout (−/−) and wild-type (+/+) mice is detected (mean ± SEM, n = 20). (B) In the 52°C water tail-immersion test, intrathecal (i.t.) injection of deltorphin I (Delt) causes an increase in tail-flick latency in a dose- and time-dependent manner in wild-type mice but not in PPT-A knockout mice (mean ± SEM, n = 8). A maximum score is assigned (100%) to animals not responding within 10 s. M.P.E., maximum possible effect. *p < 0.05 and **p < 0.01 versus the effect of the same dose in wild-type mice. (C) Intrathecal injection of morphine and baclofen (30 min) causes an increase in tail-flick latency in a dose-dependent manner in both wild-type and PPT-A knockout mice (mean ± SEM, n = 8) in the 52°C water tail-immersion test. In PPT-A knockout mice, a lower dose of morphine is needed to produce antinociception equivalent to that in wild-type mice. *p < 0.05 and **p < 0.01 versus the effect of the same dose in wild-type mice. (D) In wild-type mice, morphine tolerance is demonstrated by the loss of the analgesic response to daily subcutaneous administration (s.c.) of 5 mg morphine/kg within 5 days. The PPT-A knockout mice (mean ± SEM, n = 20) do not develop the tolerance following the same treatment of morphine for 10 days. Analgesia is tested 30 min after morphine administration with tail-flick assay. Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions

Figure 8 Proposed Model for Sorting of DOR Protachykinin is aggregated in the lumen of the trans-Golgi network and sorted into LDCVs. Interaction of the third luminal (extracellular) domain (green region) of DOR with the substance P domain of protachykinin leads to sorting DOR into LDCVs. A limited number of DORs are transported in the constitutive pathway. The stimulation triggers Ca2+ influx that induces insertion of LDCV-associated DORs (Bao et al., 2003). Cell 2005 122, 619-631DOI: (10.1016/j.cell.2005.06.010) Copyright © 2005 Elsevier Inc. Terms and Conditions